History of analytic philosophy

Language and rule following

For the logical atomists, language was conceived as having a certain necessary and fairly simple underlying structure, which was expressible in terms of symbolic logic. In other words, the underlying structure of language is reflected in the logical rules that govern the construction of molecular propositions out of atomic ones. The later Wittgenstein, however, rejected this assumption. Language, he now thought, is like an instrument that can be used for an indefinite number of purposes. Hence, any effort to codify its operation in some small set of rules would be like supposing that screwdrivers, for example, can be used only to drive screws and not also to open jars or to jimmy windows. Language is a human institution that is bound only by what its speakers consider to be correct or incorrect. And that, in turn, is not really a matter for a priori theories to consider.

The conception of language as a logical calculus leads naturally to the idea that meaning is a kind of naming, or referring. Prior to Wittgenstein’s work in the 1930s and afterward, many philosophers, especially those influenced by logical atomism, held that, for at least a large class of terms—proper names and noun phrases, for example—the meaning of the term is just the particular entity in the world that it names. This view, according to Wittgenstein, ignores the huge variety of ways in which words can be meaningful, and more generally it assumes that meaning is fixed independently of the ways in which language is used and the activities in which it is embedded. Wittgenstein’s later conception of meaning is expressed in his well-known observation (which, however, he did not make without qualification) that “the meaning of a word is its use in the language.”

The conception of language as a logical calculus was misleading in other respects as well. It suggested, for example, that in learning a language one learns the rules first and then goes on to apply them in speaking and understanding. But, in fact, one does not first learn the rules and then use the language; indeed, prior to learning the language, one would not know what to do with rules. Mathematics and logic are, in this sense, bad models for language, because they aim at setting out beforehand the rules and principles that are subsequently to be used. The “rules” that one might plausibly discern in the language that one speaks are not, as rules, already there, in a ghostly way, guiding what one says; they are either generalizations from the finite data of what is counted as correct or incorrect, or they are rules that, as Wittgenstein metaphorically expressed it, one puts away in the archives—one adopts the rule, but only after the fact.

The notion of following a rule was wrongly analyzed in many classical views about language, according to Wittgenstein. He cast irrevocable doubt on the prevalent theory—typified best, perhaps, in John Locke’s Essay Concerning Human Understanding (1690)—that to use an expression meaningfully is to have in one’s mind a standard or a rule for applying it correctly. Against this theme, Wittgenstein’s point was that a rule by itself is dead—it is like a ruler in the hands of someone who does not know how to use it, a mere stick of wood. Rules cannot compel or even guide a person unless he knows how to use them, and the same is true about mental images, which have often been thought to provide the standard for using linguistic expressions. But if rules themselves do not give life to words but require a similar explanation for what gives them life—if there must be, in effect, rules that tell one how to apply the rules—then there is a useless regress and no philosophical or explanatory value in the assumption of an apparatus of internal rules and standards.